Reliability of Accuracy-Based Calibration in Quantifying Systematic Errors of Static LiDAR

Abstract

To preserve the quality of the final spatial information, procedure of data quality assurance has become a crucial phase in geodetic jargon. Equipped with the complexity and able to acquire high precision three-dimensional data, terrestrial laser scanner (TLS) also not exempt from the calibration routine.  Due to the luxurious requirement of component calibration, most of the TLS researchers focused on the system calibration. However, in the domain of TLS system calibration, researchers have primarily focused on precision-based techniques that do not require reference points. Disregarding the cumbersome of establishing the reference points at the calibration site, the potential of accuracy-based approaches in TLS self-calibration in enhancing calibration data quality remains underexplored. This is particularly relevant as incorporating known reference points in other optical sensor calibration such as photogrammetric and laser trackers have significantly enhanced the derivation of calibration parameters (CPs). Building on these advancements, this study explores the novel application of accuracy-based approaches in TLS self-calibration using two (2) types of constraints: i) point-based and; ii) line-based. To perform constraints adjustment, the reference points were established using close-range photogrammetry (CRP) approach. Statistical analyses were performed between CPs derived from both point-based and line-based constraints and those obtained from the precision-based technique to assess the significance of these accuracy-based approaches in TLS self-calibration. The evaluation focused on four (4) critical components: i) the derivation of CPs, ii) correlation coefficient, iii) standard deviation, and; iv) residuals. Findings indicate that while point-based constraint showed promising accuracy enhancements, statistical analyses revealed significant deviations from benchmark values. This was further supported by analyses of standard deviation and residuals, highlighting notable disparities with the precision-based method. Conversely, results for line-based constraint closely aligned with the benchmark, achieving an accuracy improvement of up to 60%. To conclude, TLS self-calibration can effectively adopt an accuracy-based approach, but it is primarily limited to line-based constraint.